690.14 (C) (1) Photovoltaic Disconnecting Means

Status
Not open for further replies.
We recently had an inspector fail a Residential PV installation

His corrective action required was
  • Provide a photovoltaic disconnecting mean as required by NEC 690.14 (C) (1)


We routed the DC conductors into the attic space of the residence, then out through the soffit, and into the attached garage where the main distribution panel and inverter is located. The DC conductors are in metal conduit as soon as they enter the residence. The DC disconnecting means is the SolarEdge inverter integrated DC disconnect.


My response in writing to the inspector was as follows:
"There is an exception in 690.14 (C)(1) which I believe applies in this case. The exception allows the disconnecting means to to be remote from the point of entry so long as the installation complies with 690.31(E). 690.31 (E) requires the PV Source circuits to be contained in metal raceways, MC cable, or metal closures where it enters the dwelling. My interpretation has always been that we can take the PV source circuits directly from the array to the inverter DC disconnect as long as we comply with 690.31 (E). Can you look at this one again and let me know your thoughts."

I spoke with the inspector yesterday to follow up and he is adamant that we need a DC disconnect at the first readily accessible location inside the residence. His suggestion is that this would be in the upstairs hallway of the residence. Alternately he suggests we can run the conduit over the roof top and then down the side of the building and finally into the SolarEdge inverter DC disconnect. For residential installs we typically avoid conduits on the rootop where we can for aesthetic reasons.

I have attached a sketch to add some clarity. I would be interested in your thoughts on this.
 

Attachments

  • DC Source Conduit Run.pdf
    59.5 KB · Views: 8

SolarPro

Senior Member
Location
Austin, TX
Your interpretation is correct. You can run dc PV circuits within a building, provided they are in metal conduit. The exception to 690.14(C)(1) is very explicit about the fact that you do not need a readily accessible disconnect at the point of penetration. It says, if you comply with 690.31(E) [use metal conduit, in other words] you "shall be permitted to have the disconnecting means located remote from the point of entry of the system conductors."

This is something that AHJs often get wrong or decide to implement requirements that are above Code. And it really gets my goat, to be honest. No customer wants exposed conduit cluttering their roof or building facade if it is avoidable. The Code provides allowances for minimizing the visual impact of these conduit runs. AHJs should enforce the Code accordingly (especially in the era of rapid shutdown and dc arc fault detection.)
 

ggunn

PE (Electrical), NABCEP certified
Location
Austin, TX, USA
Occupation
Electrical Engineer - Photovoltaic Systems
Your interpretation is correct. You can run dc PV circuits within a building, provided they are in metal conduit. The exception to 690.14(C)(1) is very explicit about the fact that you do not need a readily accessible disconnect at the point of penetration. It says, if you comply with 690.31(E) [use metal conduit, in other words] you "shall be permitted to have the disconnecting means located remote from the point of entry of the system conductors."

This is something that AHJs often get wrong or decide to implement requirements that are above Code. And it really gets my goat, to be honest. No customer wants exposed conduit cluttering their roof or building facade if it is avoidable. The Code provides allowances for minimizing the visual impact of these conduit runs. AHJs should enforce the Code accordingly (especially in the era of rapid shutdown and dc arc fault detection.)

As you probably know, Austin Energy is one of those AHJ's that require a DC disco on the roof if the DC penetrates the building envelope.
 

SolarPro

Senior Member
Location
Austin, TX
Yes.
Apparently, it's a legacy of the fact that they adopted some specialty "solar code" from the early 00s. That code presumably ganked language from an early version of the NEC (probably, NEC 2002) that did require a dc disco immediately upon penetrating a building.
 

Carultch

Senior Member
Location
Massachusetts
Yes.
Apparently, it's a legacy of the fact that they adopted some specialty "solar code" from the early 00s. That code presumably ganked language from an early version of the NEC (probably, NEC 2002) that did require a dc disco immediately upon penetrating a building.

What if the DC is de-energized without a standard disconnect? Like it is in a SolarEdge system, where the optimizers default to 1 Volt per optimizer "safe DC" for 30V or less, when any DC disconnect anywhere is opened.
 

jaggedben

Senior Member
Location
Northern California
Occupation
Solar and Energy Storage Installer
Oftentimes rooftop DC disconnects are local fire department ammendment. But that doesn't seem to the be the case with our unfortunate OP whose dealing with a stubborn, unreasonable AHJ.
 

SolarPro

Senior Member
Location
Austin, TX
What if the DC is de-energized without a standard disconnect? Like it is in a SolarEdge system, where the optimizers default to 1 Volt per optimizer "safe DC" for 30V or less, when any DC disconnect anywhere is opened.

My point, exactly. Austin is finally about to adopt rapid shutdown. And any system that meets rapid shutdown requirements shouldn't need an additional dc disconnect, not even in a world where they don't adhere to the Code as it is written today.
 

SolarPro

Senior Member
Location
Austin, TX
We're about to put PV on a new "accessory dwelling unit" (garage apartment) on our alley. So it will be interesting to work through this issue with our local AHJ. The last thing I want decorating our very intentionally thought out building is a Square D disconnect that no one will ever use or have to use. Ideally, the adoption of rapid shutdown will put this to bed once and for all. :happyyes:
 

shortcircuit2

Senior Member
Location
South of Bawstin
First we have language in 690.14(C)(1) that says..."nearest the point of entry" and then we have language 690.31(E) that says..."first readily accessible location"

I disagree that this is explicit.

So the Code is saying...put it in metal raceway you can blow by the first readily accessible location and put the DC disconnect where ever you want to. (Except the bathroom which is where all these exceptions belong)
 

shortcircuit2

Senior Member
Location
South of Bawstin
What if the DC is de-energized without a standard disconnect? Like it is in a SolarEdge system, where the optimizers default to 1 Volt per optimizer "safe DC" for 30V or less, when any DC disconnect anywhere is opened.

Except if they malfunction.

How does a DC optimizer qualify as a disconnect per 690.17? Its not a manual operable switch or circuit breaker? It doesn't plainly indicate open or closed.
 

Carultch

Senior Member
Location
Massachusetts
Except if they malfunction.

How does a DC optimizer qualify as a disconnect per 690.17? Its not a manual operable switch or circuit breaker? It doesn't plainly indicate open or closed.

Because it is in series with another DC disconnect that qualifies as a disconnect per 690.17. And that other DC disconnect will comply with the location rules, other than that it isn't immediately adjacent to the array.

What the optimizers do, that doesn't happen in conventional PV circuits, is that the optimizer defaults to 1 volt on the output terminals, when the circuit is open or there is otherwise a problem with the circuit. If one optimizer malfunctions, worst case scenario, string voltage = the voltage of one module, plus the 1 volt per optimizers of the remaining functional ones. I'm not familiar with the way that they would malfunction internally, or even how they function internally, so I'm not sure if this would actually be less.

From looking at DC-DC converter schematics, it looks like there may be a transformer inside, and circuitry similar to 555 timers, that generates the transients needed to get power through it, and filters that convert it back to DC. If the DC input circuits are transformer-isolated from the DC output circuits, then don't expect voltage on the output terminals of a malfunctioning optimizer. Because steady DC voltage doesn't pass through a transformer.
http://www.eleccircuit.com/wp-conte...igh-voltage-low-current-dc-to-dc-inverter.jpg
 

SolarPro

Senior Member
Location
Austin, TX
First we have language in 690.14(C)(1) that says..."nearest the point of entry" and then we have language 690.31(E) that says..."first readily accessible location"

I disagree that this is explicit.

So the Code is saying...put it in metal raceway you can blow by the first readily accessible location and put the DC disconnect where ever you want to. (Except the bathroom which is where all these exceptions belong)

In order to deconstruct 690.14(C)(1), it helps to start from the beginning. The PV system dc disconnecting means "shall be installed at a readily accessible location either on the outside of the building or inside nearest the point of entrance of the system conductors." The either/or language is important in order to understand how to apply the Exception, which refers forward to 690.31(E).

The explanatory text in the Handbook clarifies that:

These requirement generally prohibit long runs of PV source and output circuits inside a building before reaching the required PV disconnect. A short conductor run through a wall at the point of first penetration to meet a disconnect mounted inside the building is allowed. Section 690.31(E) permits these circuits to be run inside the building when installed in metal conduit from the point of entrance to the system disconnection means.

In other words, you could, in theory, install PV source or output circuit conductors through a wall in PVC if you were going directly to a readily accessible dc disconnect on the other side of the wall. If you want to route these conductors through the building shell—say, from a 2nd story south-facing roof to an exterior disconnect alongside a load panel on the north side of the building—you can do so by following 690.31(E), as allowed by the Exception to 690.14(C)(1). [It is worth mentioning that this language moves to 690.13(A) in NEC 2014.] This language requires that the uncontrolled DC conductors are in metal conduit or Type MC Cable from the point of first penetration to wherever the PV disconnect is located. If the PV power circuits are in metal conduit or MC cable, you "shall be permitted to have the disconnecting means located remote from the point of entry of the system conductors."

This logic is clear if you read Mike Holt's Illustrated Guide to Understanding NEC Requirements for Solar Photovoltaic Systems. It is also clear if you go back and read the substantiating comments associated with the Exception to NEC 690.14 in the cyclical Report on Proposals. For example [from 2004 process leading to NEC 2005]:

Recommendation: Add a new third paragraph:

Photovoltaic systems having disconnecting means not located in compliance with the requirements of this section with respect to
nearness to the point of entry of the system conductors shall comply with 690.31(F).

Substantiation:

There are photovoltaic systems where it is physically impossible or impractical to locate a readily accessible disconnecting means at or
near the point of first penetration of the conductors. This proposal refers these installations to a proposed new Section 690.31(F) that
requires metal raceways and metal enclosures to be used for such circuits.

Later:

Recommendation:
Add the following new Section: 690.31(F) Direct Current Photovoltaic Source and Output Circuits Inside Building. Where direct current photovoltaic source or output circuits or the alternating current outputs of a utility-interactive inverter from building integrated or other photovoltaic system are run inside a building or structure, they shall be contained in metallic raceways or metallic enclosures from the point of penetration of the
surface of the building or structure to the first readily accessible disconnecting means. The disconnecting means shall comply with
690.14(A) through (C).

Substantiation:
Building integrated photovoltaic systems (defined in a new definition in 690.2) may have multiple penetration points in the surface of
a building or structure. For even a small residential system, the penetrations may number in the hundreds. It is not possible to install an
accessible disconnect at each point of penetration. Keeping these dc PV source and output circuits or other dc source and output circuits
from roof-mounted PV arrays in metallic raceways and or metallic enclosures from the point of penetration to the first readily accessible
disconnect will meet the intent of the NEC for electrical shock and fire hazard safety for these circuits.
Routing the ac outputs of
remotely located utility-interactive inverters in metallic raceways or enclosures will also increase the safety of these circuits. The
metallic enclosures may be needed in some systems to hold blocking diodes, overcurrent devices, and other equipment allowed in these
circuits before the PV disconnect by 690.14(B). Metallic raceways and enclosures provide greater mechanical protection and fire hazard
reduction than

In other words, the Code-Making Panel specifically added the requirement for metal conduit as a means of allowing installers to run uncontrolled dc PV conductors through a building shell to a conveniently located PV system disconnecting means.
 

shortcircuit2

Senior Member
Location
South of Bawstin
But the Solaredge system disconnect does't disconnect the DC Optimizer circuit conductors from its source of supply on the array side. It satisfies Rapid Shutdown, but it doesn't meet the requirements of disconnecting means.

In the link to the schematic, the DC negative is grounded. How is the Optimizer considered an ungrounded system if there is connection to ground?
 

SolarPro

Senior Member
Location
Austin, TX
In the case of the SolarEdge system specifically, you have an inverter-integrated dc disconnect below the inverter. It's a separate box from the inverter, but connects to the inverter.

Not sure what other schematic you are looking at, but the Solar Edge system complies with any Code requirements you can throw at it. In a perfect work, I'd like to use SolarEdge for my home system. But I'm not sure that SunPower will let their integrator partners sell SunPower modules with a SolarEdge power conversion system.
 

Attachments

  • PastedGraphic-1.jpg
    PastedGraphic-1.jpg
    21.9 KB · Views: 2
Last edited:

shortcircuit2

Senior Member
Location
South of Bawstin
In order to deconstruct 690.14(C)(1), it helps to start from the beginning. The PV system dc disconnecting means "shall be installed at a readily accessible location either on the outside of the building or inside nearest the point of entrance of the system conductors." The either/or language is important in order to understand how to apply the Exception, which refers forward to 690.31(E).

The explanatory text in the Handbook clarifies that:



In other words, you could, in theory, install PV source or output circuit conductors through a wall in PVC if you were going directly to a readily accessible dc disconnect on the other side of the wall. If you want to route these conductors through the building shell—say, from a 2nd story south-facing roof to an exterior disconnect alongside a load panel on the north side of the building—you can do so by following 690.31(E), as allowed by the Exception to 690.14(C)(1). [It is worth mentioning that this language moves to 690.13(A) in NEC 2014.] This language requires that the uncontrolled DC conductors are in metal conduit or Type MC Cable from the point of first penetration to wherever the PV disconnect is located. If the PV power circuits are in metal conduit or MC cable, you "shall be permitted to have the disconnecting means located remote from the point of entry of the system conductors."

This logic is clear if you read Mike Holt's Illustrated Guide to Understanding NEC Requirements for Solar Photovoltaic Systems. It is also clear if you go back and read the substantiating comments associated with the Exception to NEC 690.14 in the cyclical Report on Proposals. For example [from 2004 process leading to NEC 2005]:



Later:



In other words, the Code-Making Panel specifically added the requirement for metal conduit as a means of allowing installers to run uncontrolled dc PV conductors through a building shell to a conveniently located PV system disconnecting means.

Thank you for the detailed response Solarpro.

But I hope you can see how the language can be interpreted to mean that the disconnect should be at the first readily accessible location. There is no language that says conveniently located where you want if in metal raceway.

You are telling an industry that has been taught to have a disconnecting means for power to a building that "shall be installed at a readily accessible location either on the outside of the building or inside nearest the point of entrance of the system conductors."​ to think otherwise. See the confusion it has created? Now the we are saying put it in metal raceway and conveniently locate it were you want?

We could be talking about circuits up to 1000-volts here. I can't see the safety in this practice.
 

shortcircuit2

Senior Member
Location
South of Bawstin
Because it is in series with another DC disconnect that qualifies as a disconnect per 690.17. And that other DC disconnect will comply with the location rules, other than that it isn't immediately adjacent to the array.

What the optimizers do, that doesn't happen in conventional PV circuits, is that the optimizer defaults to 1 volt on the output terminals, when the circuit is open or there is otherwise a problem with the circuit. If one optimizer malfunctions, worst case scenario, string voltage = the voltage of one module, plus the 1 volt per optimizers of the remaining functional ones. I'm not familiar with the way that they would malfunction internally, or even how they function internally, so I'm not sure if this would actually be less.

From looking at DC-DC converter schematics, it looks like there may be a transformer inside, and circuitry similar to 555 timers, that generates the transients needed to get power through it, and filters that convert it back to DC. If the DC input circuits are transformer-isolated from the DC output circuits, then don't expect voltage on the output terminals of a malfunctioning optimizer. Because steady DC voltage doesn't pass through a transformer.
http://www.eleccircuit.com/wp-conte...igh-voltage-low-current-dc-to-dc-inverter.jpg


The schematic here in the link above...
 

SolarPro

Senior Member
Location
Austin, TX
Where I say "conveniently located," feel free to read "readily accessible location, located remote from the point of entry."

Like I said, this is an often misinterpreted section of the Code. And I agree that it is ripe for interpretation. Having been in the industry while these Code changes were being implemented—and having attended trainings over the years presented by members of the Code-Making Panel—I'd like to think I have some insight into the intent of the Code language. But I'm not an AHJ. And at the end of the day, that's the only opinion/interpretation that matters.
 

shortcircuit2

Senior Member
Location
South of Bawstin
Where I say "conveniently located," feel free to read "readily accessible location, located remote from the point of entry."

Like I said, this is an often misinterpreted section of the Code. And I agree that it is ripe for interpretation. Having been in the industry while these Code changes were being implemented—and having attended trainings over the years presented by members of the Code-Making Panel—I'd like to think I have some insight into the intent of the Code language. But I'm not an AHJ. And at the end of the day, that's the only opinion/interpretation that matters.

Once again thank you for your valuable input here.

It doesn't read "readily accessible location" It reads "first readily accessible location"...there lies the confusion.

You and the code panel know what you were thinking...but the text of the Code and an AHJ erring on the side of safety will see it differently as in the OP case.
 

Carultch

Senior Member
Location
Massachusetts
But the Solaredge system disconnect does't disconnect the DC Optimizer circuit conductors from its source of supply on the array side. It satisfies Rapid Shutdown, but it doesn't meet the requirements of disconnecting means.

In the link to the schematic, the DC negative is grounded. How is the Optimizer considered an ungrounded system if there is connection to ground?

That's not necessarily the schematic for a SolarEdge DC optimizer. That is a schematic for simply one example of a DC to DC converter. An example unrelated to solar.

The optimizer system is considered an ungrounded system because both polarities are intentionally ungrounded conductors, built to work at equal and opposite voltages for the inverter to build its AC circuit.
 
Status
Not open for further replies.
Top